“This study highlights the influence of breast density on the diagnostic accuracy of screening DBT, reinforcing the importance of mandatory breast density notification and personalized screening regimens to mitigate the risks associated with density.” noted lead author Ariel S. Kniss, M.D., Ph.D., a radiology fellow at the Massachusetts General Hospital in Boston, and colleagues.
In multivariable analysis, the researchers noted that increasing breast density was correlated with higher abnormal interpretation rates (AIRs) and lower specificity rates with one exception.
Three Key Takeaways
• Marked reduction in DBT sensitivity with extreme breast density. Screening digital breast tomosynthesis (DBT) sensitivity was over 30 percent lower in women with extremely dense breasts compared with those with entirely fatty breasts (61.8 percent vs. 92.8 percent), confirming that density remains a major limitation for cancer detection even with DBT.
• Higher false-negative risk in dense breasts. Women with extremely dense and heterogeneously dense breasts had substantially higher false-negative rates (2.3 and 1.2 per 1,000 DBT exams respectively) than women with fatty or scattered fibroglandular tissue, reinforcing the need for supplemental or personalized screening strategies in these populations.
• Complex effects of density on recall and specificity. Increasing breast density was generally associated with higher abnormal interpretation rates (AIRs) and lower specificity. However, extremely dense breasts showed lower AIR and higher specificity than heterogeneously dense breasts, likely reflecting small sample size or nuanced differences in lesion conspicuity across DBT slices—highlighting the challenges of interpreting DBT in dense tissue, the importance of density notification and tailored screening.
The study authors found that women with extremely dense breasts (category D) had a lower AIR (0.78 adjusted odds ratio (aOR)) and higher specificity (1.28 aOR) in contrast to women with heterogeneously dense breasts (category C). This finding may result from statistical underpowering due to the lower number of women with category D presentations in the cohort or minimal differences with conspicuity across multiple imaging slices for women with dense fibroglandular tissue, according to the researchers.
“These observations underscore the complexity of interpreting screening DBT examinations with dense tissue and highlight the need for a nuanced understanding of how different density categories affect screening performance,” pointed out Kniss and colleagues.
(Editor’s note: For related content, see “Mammography Study: Multi-Stage Use of AI for DBT Exams Yields 21 Percent Increase in Breast Cancer Detection,” “What a DBT Screening Study Reveals About False Positives with AI and Radiologist Assessments” and “Digital Breast Tomosynthesis Study Assesses Impact of Architectural Distortion on Malignancy Rates.”)
Beyond the inherent limitations of a single-center retrospective study, the authors acknowledged the subjective nature of breast density classification without the use of automated software.
